Prince Canuma

Canuma’s current GitHub biography describes work on machine-learning operations, language models, vision-language models and audio, and explicitly marks his Arcee AI research-engineer role as former. His linked professional profile names Neywa Labs and Wroclaw University of Science and Technology, but its public education details do not establish a degree or dates.[1][5]

A shared path for images, video and text

MLX-VLM supports inference and fine-tuning of vision-language and multimodal models on Macs using Apple’s MLX framework. It supplies command-line, Python and server interfaces, along with model-specific examples and memory-saving options. The contribution is an implementation and integration layer for many architectures, rather than authorship of all the underlying models.[2]

Audio libraries for Python and Swift applications

MLX-Audio covers speech generation, transcription, speech-to-speech processing and music models. Its Swift counterpart separates shared utilities, codecs, speech generation, recognition and interface components into modules, letting an app import only the pieces it needs. That connects model experimentation on a Mac to applications built for Apple devices.[3][4]

From libraries to a local application

Neywa’s official site identifies MLX-VLM and MLX-Audio as its libraries and presents Nativ as an app for running models and connecting developer tools through standard APIs. This provides current organizational context for Canuma’s open-source work without implying that every library feature or company product was written by him alone.[5][6]